pySPL.py

from lxml import etree
from datetime import date

#http://www.accessdata.fda.gov/spl/stylesheet/spl-common.xsl
namespaces={"v3":"urn:hl7-org:v3",}

def normalize_date(date_string):
	year = int(date_string[0:4])
	month = int(date_string[4:6])
	day = int(date_string[6:8])
	return date(year,month,day).strftime("%b %d, %Y")

class DrugLabel(object):
	"""represents a Drug Label in the SPL format.
	takes one argument, spl_label, which can be either an url or a file path"""
	
	def __init__(self, spl_label):
		self.label_data = spl_label
		self.xml = etree.parse(spl_label)
	
	def actives(self):
		"""returns a list of active compounds"""
		#here converting to a set removes duplicates
		return sorted(list(set(active.text for active in self.xml.xpath("//v3:ingredientSubstance/v3:activeMoiety/v3:activeMoiety/v3:name",namespaces=namespaces))))
	actives.label = "active cmpds"

	def start_date(self):
		"""returns start marketing date as a strftime formatted python date object"""
		date_string = self.xml.xpath("//v3:subjectOf/v3:marketingAct/v3:effectiveTime/v3:low/@value",namespaces=namespaces)[0]
		return normalize_date(date_string)
	start_date.label = "marketing start date"

	# don't bother - it's None for all labels
	# end_date.label = "marketing end date"
	# def end_date(self):
	# 	"""returns end marketing date as a strftime formatted python date object or the string "None if not defined
	# 	refers to the expiration date of the last lot released to the market
	# 	(from http://spl-work-group.wikispaces.com/file/view/creating_otc_sp_documentsl.pdf)"""
	# 	try:
	# 		date_string = self.xml.xpath("//v3:subjectOf/v3:marketingAct/v3:effectiveTime/v3:high/@value",namespaces=namespaces)[0]
	# 		return normalize_date(datestring)
	# 	except:
	# 		return "None"
	
	def marketing_category(self):
		"""returns the marketing category"""
		return self.xml.xpath("//v3:subjectOf/v3:approval/v3:code/@displayName",namespaces=namespaces)[0]
	marketing_category.label = "marketing category"
	 
	def revision_date(self):
		"""returns label revision date"""
		date_string = self.xml.xpath("/v3:document/v3:effectiveTime/@value",namespaces=namespaces)[0]
		return normalize_date(date_string)
	revision_date.label = "revision date"
	
	def label_type(self):
		"""returns the drug label type, typically 'HUMAN OTC DRUG LABEL' or 'HUMAN PRESCRIPTION DRUG LABEL' """
		return self.xml.xpath("//v3:code/@displayName",namespaces=namespaces)[0]
	label_type.label = "label type"
	

	def ndc(self):
		"""returns the drug's NDC number"""
		#this xpath query is NOT from the SPL xsl file
		return self.xml.xpath("//v3:manufacturedProduct/v3:manufacturedProduct/v3:code/@code",namespaces=namespaces)[0]
	ndc.label = "ndc"
	
	def name(self):
		"""returns the drug's name"""
		return self.xml.xpath("//v3:manufacturedProduct/v3:manufacturedProduct/v3:name",namespaces=namespaces)[0].text.replace("\t","").replace("\n","")
	name.label = "name"

	def distributor(self):
		"""returns the drug's distributor"""
		return self.xml.xpath("//v3:author/v3:assignedEntity/v3:representedOrganization/v3:name",namespaces=namespaces)[0].text
	distributor.label = "distributor"
	
	def dosage_form(self):
		"""returns the drug's dosage form"""
		return self.xml.xpath("//v3:manufacturedProduct/v3:manufacturedProduct/v3:formCode/@displayName",namespaces=namespaces)[0]
	dosage_form.label = "dosage form"
	
	#just a helper function for the next two functions, so no label
	def _get_word_list(self, word):
		"""returns a list of etree instances of all occurances of 'word','Word' or 'WORD' """
		word = str(word)
		#this query also NOT from the SPL xsl file
		query = "//*[text()[contains(.,'%s') or contains(.,'%s') or contains(.,'%s')]]" %(word.lower(),word.upper(),word.capitalize())
		return self.xml.xpath(query,namespaces=namespaces)

	def test_word(self, word):
		if self._get_word_list(word): return 1
		else: return 0
	#TODO test_word.label = "%s?" %self.test_word.word
	
	def get_word_section(self, word):
		#finds the first ancestor section and returns section/code/@displayName
		word_section_list = []
		for word in self._get_word_list(word):
			word_section_list.extend(word.xpath("ancestor::v3:section[1]/v3:code/@displayName",namespaces=namespaces))
		if not word_section_list:
			return "n/a"
		return list(set(word_section_list))
	
	def get_word_time(self, word):
		"""
		returns the LATEST effectiveTime/@date for all of the instances of "word" mentioned
		"""
		word_time_list = []
		for word in self._get_word_list(word):
			word_time_list += word.xpath("ancestor::v3:section/v3:effectiveTime/@value",namespaces=namespaces)
		if not word_time_list:
			return "n/a"
		try:
			return max(list(set(normalize_date(date) for date in word_time_list)))
		except:
			return max(list(normalize_date(date) for date in word_time_list))

	def build_url(self):
		"""helper function that builds and returns the accessdata.fda.gov URL given the XML file name/directory"""
		#maybe won't work on windows because slash direction?
		uuid = self.label_data.split("/")[-1].split(".")[0]
		return "http://www.accessdata.fda.gov/spl/data/%s/%s.xml" %(uuid,uuid)
	build_url.label = "url"